Generating True/False Quizzes for STEM With AI
In 1947, General Electric engineer Lawrence D. Miles formalized a wartime materials-shortage workaround into a repeatable method: instead of asking which design option was "best," his teams asked which one delivered the required function at the lowest cost, given a stated set of constraints (Miles, 1961). That discipline, value engineering, is the direct ancestor of today's engineering-design trade-off item.
It also previews a problem this guide has to work through honestly. Change which constraint you weight, and the "right" answer changes with it. A true/false claim only ever gets one verdict.
Quick Answer: STEM's true/false territory is narrower than biology's, and shaped differently. A full engineering trade-off resists true/false structurally, since it requires comparing options — something one bare claim can't do. A compute item resists it too, since a claim built around one number functions like a masked multiple-choice option. True/false's real niche sits one layer down: a stated relationship (does doubling one quantity double or quadruple another) or a single constraint check pulled from a trade-off, not the full comparison or computation.
How to Create STEM Quizzes With AI already worked out what makes a quiz item genuinely STEM rather than a single subject wearing an acronym: an item that fails unless a student combines a computation with a science or engineering concept, or evaluates a design decision against explicit constraints. This guide, alongside How to Create Spanish Quizzes With AI in this site's subject-specific quiz series, asks a narrower question about that same territory: which of it, if any, survives being rewritten as a bare true/false claim?
This guide covers:
- Why most single-subject STEM claims already have a better home in existing guidance
- Why a full engineering trade-off decision resists a clean true/false verdict, and what narrower claim survives instead
- Why a cross-disciplinary compute item resists it too, unless you test the relationship instead of the number
- The Single-Claim Test: a check for whether a STEM claim commits to exactly one checkable proposition
- A step-by-step AI workflow, a tools comparison, and the pitfalls that undercut an otherwise solid item
What "STEM True/False" Actually Means
A true/false claim about a STEM topic is only genuinely STEM-shaped if verifying it requires both a computation and a science or engineering concept — the same bar the STEM quiz guide set for multiple-choice items. A claim that's really about one subject, physics, chemistry, or a single engineering fact, is a single-subject true/false claim with a STEM label attached, not new territory.
Three Requests, One Genuine Answer
| If You're Asking For... | Where It Belongs | Why |
|---|---|---|
| A true/false claim testing one physics, chemistry, or biology fact | A single-subject guide, or the true/false craft guide | One discipline decides the verdict; the STEM label doesn't change the item-writing mechanics |
| A claim testing whether a design decision satisfies stated constraints | This guide, with a caveat covered below | The decision itself needs two disciplines to evaluate |
| A claim testing whether a computed value or a scaling relationship holds | This guide | The verdict depends on both a formula and doing math with it |
Single-subject physics claims have their own documented misconception research — the Force Concept Inventory catalogued Aristotelian errors like "a constant force is needed to keep an object moving at constant speed" across thousands of students (Hestenes, Wells, & Swackhamer, 1992). Those are real, testable, cleanly binary claims. They aren't cross-disciplinary, so the true/false craft guide already covers writing that kind of claim well.
The Two Item Types This Guide Has to Test
What's left, after routing away single-subject claims, are the two item types the STEM quiz guide identified as genuinely cross-disciplinary: a compute item, where a concept determines which calculation to run, and a trade-off item, where a design decision gets evaluated against explicit, sometimes competing constraints. Both were built and validated as multiple-choice items, not with a true/false verdict in mind, and reformatting each one exposes a different problem.
Why This Isn't the Same Question Biology Answered
The biology true/false guide found misconceptions that hold the same verdict across every case they describe — a population-of-cases uniformity problem. STEM's defining items aren't built around a population of cases at all. They're built around a comparison (the trade-off) and a computation (the compute item) — a structurally different reason to resist a single verdict, not a milder version of biology's problem.
Why a Trade-Off Decision Resists a Clean Verdict
An engineering-design trade-off item, as the STEM quiz guide builds it, asks a student to compare two or more options against stated constraints and identify which one satisfies all of them. That comparison is the whole point of the exercise, and a true/false claim can only ever evaluate one proposition at a time, never a comparison across several.
A Worked Example
Say a grade 6 class is designing an insulated lunch-bag liner and comparing three materials against two constraints: cost under $2 per square foot, and an insulation rating (R-value) of at least 3.
| Material | Cost per Sq Ft | R-Value | Meets Both Constraints? |
|---|---|---|---|
| Foam board | $1.50 | 4 | Yes |
| Bubble wrap | $1.00 | 2 | No — fails R-value |
| Wool felt | $2.50 | 5 | No — fails cost |
Only foam board clears both bars — a clean, gradable multiple-choice item: compare three rows, rule out two, name the winner.
Why "Foam Board Is Correct" Still Isn't a Good True/False Claim
Rewrite that conclusion as "foam board is the correct insulation material for this design" and ask true/false. A student can verify it by checking foam board alone against both constraints, never looking at bubble wrap or wool felt at all.
The comparative reasoning the trade-off item exists to test, ruling out competing options rather than just confirming one, disappears the moment the format only asks about a single named option. That's a structural mismatch no rewording fixes.
The Narrower Claim That Does Survive
One piece of that trade-off item is genuinely true/false-shaped: a single constraint check against a single option, pulled out of the full comparison.
- Doesn't survive: "Foam board is the best material for this design." (a comparison across all three options)
- Survives: "Bubble wrap's R-value of 2 meets the unit's minimum requirement of R-3." (False — one option, one stated number, one threshold)
That narrower claim tests whether a student can read a constraint correctly against a data table — a real, useful, gradable skill. It's just not the same skill as making the trade-off decision itself, and a true/false item shouldn't be mislabeled as testing the second when it only tests the first.
Why a Compute Item Resists It Too, Unless You Test the Relationship
A cross-disciplinary compute item resolves to a specific number, and a true/false claim built around a specific number behaves like one multiple-choice option wearing a disguise. That's a different failure than the trade-off item's comparison problem, but it's just as real.
The "One Option in Disguise" Problem
Take a cart with a mass of 2 kg moving at 4 m/s. Its kinetic energy is 16 joules (KE = ½mv² = 0.5 × 2 × 4² = 16). Written as a true/false claim, "a 2 kg cart moving at 4 m/s has a kinetic energy of 16 joules" is a correct, gradable statement.
It's also weak. A student can mark it true or false at even odds, without running the formula, unless the item also requires a shown calculation or one-sentence justification — the same "why" requirement the biology true/false guide recommends. Strip that requirement away, and the item measures guessing more than computing.
The Reframe: Test the Relationship, Not the Number
A stronger cross-disciplinary claim doesn't ask whether one number is correct. It asks whether a relationship holds, across every case that shares its structure, not just the one instance in front of the student.
Take that same cart. If its velocity doubles from 4 m/s to 8 m/s while its mass stays fixed, does its kinetic energy double too? KE at 8 m/s is 0.5 × 2 × 8² = 64 joules, four times the original 16, not double.
"Doubling an object's velocity while its mass stays constant also doubles its kinetic energy" is false, and it's false for every mass and starting velocity, not just this pair of numbers.
That claim needs both disciplines: the concept (kinetic energy depends on velocity squared) and the math (squaring a doubled value quadruples it, not doubles it). Removing either half breaks the claim — the same removal test the STEM quiz guide applies to compute items generally.
Not Every Scaling Claim Quadruples
| Relationship | If You Double the Input | Verdict on "Output Also Doubles" |
|---|---|---|
| Kinetic energy vs. velocity (mass fixed) | Quadruples | False |
| Voltage vs. current, fixed resistance (Ohm's Law) | Doubles | True |
| A circle's area vs. its radius | Quadruples | False |
| Force vs. acceleration, fixed mass (Newton's second law) | Doubles | True |
That variety matters. A set that's always "doubling quadruples the output" trains pattern-matching, not reasoning about the formula's actual shape.
The Single-Claim Test for STEM True/False Items
A STEM claim is only true/false-shaped if it commits to exactly one checkable proposition, not a comparison across multiple options, and not a bare number standing in for "the correct answer goes here" with no relationship or reasoning attached. Call that the Single-Claim Test.
Running the Test
- Check whether verifying the claim requires comparing two or more options. If it does, the claim belongs in a multiple-choice or ranking format instead; no rewording fixes a comparison problem.
- Check whether the claim is only a bare computed number. If so, it's gradable but weak; require a shown calculation or a one-sentence justification, or it mostly measures guessing.
- Check whether the claim states a relationship or proportion. A claim about how one quantity changes as another does passes cleanly, since it holds across every case with that structure, not one instance.
- Check whether the claim isolates one option against one constraint, pulled out of a larger trade-off. That passes too, but be precise about what it tests: constraint literacy, not the comparison itself.
A Quick Reference
| Claim Type | Passes the Single-Claim Test? | Why |
|---|---|---|
| "Option A is the best choice for this design." | No | Requires comparing every option, not just verifying one |
| "A 2 kg cart at 4 m/s has 16 J of kinetic energy." | Gradable, but weak alone | One bare number; pair with a justification requirement |
| "Doubling velocity quadruples kinetic energy when mass stays fixed." | Yes | A relationship true across every matching case |
| "Bubble wrap's R-value meets the unit's R-3 minimum." | Yes | One option, one stated constraint, checkable against given data |
STEM Claims That Pass the Test
The claims that pass the Single-Claim Test cluster into two honest categories: scaling relationships between a science quantity and a math operation, and single-constraint checks pulled from a larger design comparison. Both are narrower than the full trade-off decision or the full compute problem — that's the honest finding, not a workaround to avoid stating it.
Scaling Relationships by Grade Band
A grade 5 unit on simple machines could test: "doubling the length of a lever's effort arm doubles the force it takes to lift a fixed load." That's false; it halves the force, an inverse relationship rather than a direct one. A grade 8 physics unit could reuse the kinetic-energy claim above, or Boyle's Law: "halving a gas's volume at constant temperature doubles its pressure" — true, and inverse rather than squared.
A Grade 9 Illustration
Say you teach a grade 9 physical science class finishing a circuits unit. Alongside an Ohm's Law claim, you could ask whether doubling a wire's cross-sectional area, keeping length and material fixed, halves its resistance. It does, since resistance is inversely proportional to cross-sectional area. Pairing that with a squared or linear claim forces students to track which formula fits which relationship, rather than defaulting to whichever pattern the last item used.
Constraint Checks by Grade Band
Say you teach a grade 4 class running a classic egg-drop challenge, padding an egg against a 200-gram weight limit and a $3 budget. A true/false set could isolate one fact from the design brief: "using both cotton balls and bubble wrap keeps the total padding under 200 grams" is checkable against whatever the class actually measured, without asking a student to have already picked a winning design.
A Comparison Worth Keeping in View
None of this replaces the STEM quiz guide's multiple-choice trade-off item for testing the comparison itself. True/false's role here is narrower on purpose: a fast, frequent check on a relationship or a single constraint, sitting alongside the deeper multiple-choice item rather than replacing it.
A Step-by-Step AI Workflow for STEM True/False Items
Building a defensible STEM true/false item starts with running the Single-Claim Test before drafting, since a relationship claim, a constraint check, and a full trade-off decision each need a different format.
- Identify whether the claim is a relationship, a constraint check, or a comparison, using the Single-Claim Test before writing anything.
- If it's a comparison, stop and route it to a multiple-choice or ranking item. The STEM quiz guide already covers building that format well.
- If it's a relationship, state it as a proportional claim ("doubling X does/doesn't double Y"), not a single computed pair of numbers.
- If it's a constraint check, name one option and one stated threshold explicitly, not the full design brief.
- Ask AI to show the underlying calculation or formula it used, so you can verify a "true" relationship claim before it reaches students.
- Require a one-sentence justification on at least half the set, the same reasoning-check move the biology true/false guide recommends, since a bare relationship claim still carries 50/50 guess odds without it.
- Set your class profile's grade level and current unit, so the formulas and vocabulary match what the class has actually covered.
- Balance the true/false ratio and randomize order, the craft step the true/false speed guide covers in depth for any subject.
You could use EduGenius to hold a class profile's grade level and current unit while drafting a batch of scaling-relationship or constraint-check true/false items, then export the finished set with an automatic answer key that shows the underlying calculation, not just the verdict.
Tools for STEM True/False Quizzes
A tool built for this content needs to support an explicit relationship-versus-comparison distinction and a shown calculation, two things a generic true/false generator doesn't ask for by default.
| Tool | Distinguishes Relationships From Comparisons | Shows Underlying Calculation | Class-Context Memory | Starting Price |
|---|---|---|---|---|
| EduGenius | Only if the Single-Claim distinction is requested explicitly | Yes, if requested | Yes, via saved class profiles | 25 welcome credits free; Starter $7.99/month (500 credits); Professional $15.99/month (1,000 credits) |
| General-purpose AI chat tools | Only if explicitly requested | Only if explicitly requested | No, unless re-typed each time | Varies by provider |
| Publisher STEM item banks | Reliable within a pre-reviewed bank | Rarely shown to students | N/A | Included with curriculum adoption |
| Manual, teacher-written | Full control | Teacher's own work shown, if included | N/A | No cost |
No tool replaces working a relationship claim's math by hand before it reaches a class, the same independent check the STEM quiz guide recommends for any quantitative content.
Pro Tips for STEM True/False Quizzes
- Name the relationship in your prompt, not just the topic. "A true/false item testing whether doubling radius quadruples a circle's area" produces a sharper, more verifiable item than "a hard geometry question."
- Run the Single-Claim Test on every drafted item before it reaches students — especially anything that started life as a trade-off comparison.
- Pair a scored true/false set with a Bloom's-aligned check the week a formula or design unit is first introduced. AI True/False Quizzes Aligned to Bloom's Taxonomy covers matching item difficulty to where a unit actually stands.
- Reserve the full trade-off decision for a multiple-choice or ranking format. The STEM quiz guide's engineering-design item is still the better tool for testing the comparison itself, not just one constraint pulled from it.
What to Avoid
Rewriting a Trade-Off Winner as a Bare True/False Claim
"Option A is correct" lets a student verify one option without ever ruling out the others, so the comparison the trade-off item exists to test quietly disappears. Isolate one constraint instead, or keep the full decision as multiple-choice.
Testing a Computed Number Without a Justification Requirement
A true/false claim built around one specific number has even guess odds unless it also requires a shown calculation or a one-sentence "why." Without that, it measures guessing more than computing.
Assuming Every Doubling Quadruples the Output
Some STEM relationships are linear (doubling force doubles acceleration), some are quadratic (doubling velocity quadruples kinetic energy), and some are inverse (doubling volume halves pressure, at constant temperature). Mixing relationship shapes in a set prevents the pattern-matching shortcut.
Skipping the Independent Math Check
A confidently worded relationship claim is only as good as the algebra behind it. Plug real numbers into the formula on both sides before trusting a generated verdict, the same discipline any compute item needs.
Key Takeaways
- STEM's true/false territory is narrower than biology's, not a direct parallel. Biology supplies genuinely categorical misconceptions; STEM's two defining item types resist a bare verdict more often than they accept one.
- A full engineering trade-off decision resists true/false structurally, since verifying it requires comparing multiple options, and a true/false claim only ever evaluates one proposition (Miles, 1961).
- A single constraint check pulled from a trade-off does survive — a narrower, real skill (constraint literacy), honestly distinct from the comparison itself.
- A cross-disciplinary compute item built around one number is gradable but weak, functioning like a masked multiple-choice option unless paired with a required justification.
- A relationship claim, does doubling one quantity double, quadruple, or halve another, is STEM true/false's strongest genuine niche, since it holds across every matching case, not one instance.
- Not every scaling relationship behaves the same way. Mixing linear, quadratic, and inverse relationships in one set keeps students reasoning about the formula instead of pattern-matching the answer.
- AI can draft relationship claims and show its calculation quickly; a human still verifies the math and the constraint data before either reaches a class.
Frequently Asked Questions
Is true/false a good format for STEM quizzes?
It depends which STEM item you mean. A full engineering trade-off decision or a bare computed number both resist a clean true/false verdict. A scaling relationship or a single constraint check pulled from a larger design comparison both work well as true/false claims.
Can a true/false claim test an engineering design trade-off?
Not the full decision. Comparing multiple options against constraints structurally needs a multiple-choice or ranking format, since a true/false claim only ever evaluates one proposition. A single constraint check on one option, pulled out of that comparison, does work as true/false.
Why isn't "the answer is 16 joules" a strong true/false item?
It functions like one multiple-choice option without the other three, giving even guess odds with no computation required. A shown calculation or one-sentence justification fixes that; testing a relationship, like whether doubling velocity quadruples kinetic energy, fixes it more directly.
How is this different from the STEM quiz guide's multiple-choice items?
That guide builds compute and trade-off items as multiple-choice, where comparing options or eliminating distractors is part of the task. True/false strips that scaffolding away, which is exactly why a full trade-off decision doesn't survive the conversion: the comparison has nowhere to go inside a single claim.
Sources
- Miles, L. D. (1961). Techniques of Value Analysis and Engineering. McGraw-Hill.
- NGSS Lead States. (2013). Next Generation Science Standards: For States, By States.
- National Research Council. (2012). A Framework for K-12 Science Education: Practices, Crosscutting Concepts, and Core Ideas. National Academies Press.
- Hestenes, D., Wells, M., & Swackhamer, G. (1992). Force concept inventory. The Physics Teacher, 30(3), 141–158.
- Frisbie, D. A. (1992). The multiple true-false item format: A status review. Educational Measurement: Issues and Practice, 11(4), 21–26.
- NCTM. (2014). Principles to Actions: Ensuring Mathematical Success for All. National Council of Teachers of Mathematics.
- OECD. (2023). PISA 2022 Assessment and Analytical Framework. OECD Publishing.